Patience and achievement test results

ABSTRACT

I investigate the relation between high-stakes achievement test scores and students’ patience. I use an experimentally validated measure of time preferences to assess patience in a large sample of Dutch students in secondary school. A one SD increase in the achievement test at the age of 12 predicts 26% of an SD greater level of patience at the age of 15. The results hint at the fact that education can shape preferences.

KEYWORDS:

• Patience
• achievement test scores
• time preferences
• discount rate
• Cito test

JEL CLASSIFICATION:

• I20
• I21

Acknowledgements

I thank Bas ter Weel and three anonymous referees for feedback which helped improving the article. The data and the programmes which were used to derive the results are available upon request.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Note that in the course of the article, I will use the terms patience, impatience and time preferences interchangeably for the same underlying concept. This is the evaluation of a good in future periods compared to the same good today.

² The data set is called Onderwijs Monitor Limburg (OML). More information is provided in the online appendix.

³ Some studies also use un-incentivized measures of discount rates where individuals had to make decisions between earlier smaller amounts and delayed greater amounts (e.g. Golsteyn, Grönqvist, and Lindahl 2014; Non and Tempelaar 2014). However, for these measures, it is not clear whether individuals would give similar answers if the amounts were actually paid to them. Moffitt et al. (2011) use a battery of nine items to assess an individual’s degree of self-control. They use a rating of teachers, parents and self-reported answers of items such as fleeting attention, lacking persistence and impulsivity, to assess self-control of children between the ages of 3 and 11. This method of assessment has the advantage of reducing measurement error of the respective trait of interest since one has multiple observers for the same item. On the other hand, it is often not feasible in huge data sets to get this detailed information.

⁴ Further information on the variables is available upon request.

Additional information

Funding

The research was funded by Meteor, Maastricht.